User Evaluation to Measure the Perception of Similarity Measures in Artworks

Similarity measures do not typically capture subjective elements of perception similarity. Our research contributes an experimental methodology for validating and learning similarity computation algorithms against human perceptions in domains like Art emotions. In this paper, we explain the first experiment to check our hypotheses methodology. We have obtained promising results that differences between users profiles their artworks how combine local functions be able compute reflecting users’ perception.